Determinants of Vertical Integration: Financial Development and Contracting Costs

We study the determinants of vertical integration in a new data set of over 750,000 firms from 93 countries. We present a number of theoretical predictions on the interactions between financial development, contracting costs, and the extent of vertical integration. Consistent with these predictions, contracting costs and financial development by themselves appear to have no effect on vertical integration. However, we find greater vertical integration in countries that have both greater contracting costs and greater financial development. We also show that countries with greater contracting costs are more vertically integrated in more capital-intensive industries. CASUAL EMPIRICISM SUGGESTS THE PRESENCE OF SIGNIFICANT differences in the organization of production across countries. For example, firms are often thought to be larger and more vertically integrated in less developed countries. Khanna and Palepu (1997, 2000) provide evidence consistent with this view and suggest that this is because market and contractual relationships are more costly in less developed countries. Nevertheless, there has not been a systematic analysis of cross-country differences in vertical integration and their causes. Our primary aim in this paper is to make a first attempt at such a systematic analysis and to investigate the relationship between important institutional characteristics and vertical integration across countries. Two well-established theories offer predictions on how differences in (specific) institutional characteristics of countries should affect the internal organization of the firm in general and vertical integration in particular. First, according to the highly influential Transaction Cost Economics (TCE) theory pioneered by Williamson (1975, 1985), the internal organization of a firm is designed to improve incentives and limit agency costs. Vertical integration is perhaps the best known application of this theory. Vertical integration encourages specific investments and reduces holdup problems when markets are imperfect. According to TCE, vertical integration should therefore be more prevalent when it is harder to write long-term contracts between upstream and downstream firms. This prediction is not entirely unambiguous, however. The more sophisticated ∗Acemoglu and Johnson are at MIT, and Mitton is at Brigham Young University. We acknowledge the helpful comments of John McMillan, Rob Stambaugh, and an anonymous referee. We thank David Autor, Joseph Fan, and Nathan Nunn for providing data. We are particularly grateful to Jacques Crémer and Jianye Yan for pointing out an error in an earlier version of the paper. Acemoglu gratefully acknowledges financial support from the National Science Foundation.